CounterCall

Company: CounterCall

Self-reported basis: The analysis is based entirely on the author's own description of CounterCall, as submitted to the OpenAI 2026 hackathon on Devpost. No external verification or historical data are available.

What it appears to be: A browser-based application designed to train football referees using counterfactual reasoning — a method that requires learners to identify the decisive observable fact in a decision and then apply that principle to a modified scenario.

What changed: The author states they built this independently in Moldova, with no federation endorsement or product affiliation.

Most important open question: Does CounterCall demonstrate a viable path to scalable training for referees, or is it limited to a narrow use case?

The description states that CounterCall is a counterfactual reasoning coach for football referees. It operates by:

• Requiring learners to commit to a decision and name the decisive observable cue.
• Presenting a near-identical incident in which exactly one decision-relevant fact has changed.
• Requiring the learner to transfer the principle, not simply remember an answer.
• Separating Law, Fact, and Judgement.
• Not counting correct guessing without reasoning as mastery.

The product is described as a responsive browser application with a Node.js server and Cloudflare Worker-compatible deployment. It uses GPT-5.6 Sol through the OpenAI Responses API for personalized coaching, but the model does not affect grading.

Inference: The tool appears to be an experimental educational platform focused on reasoning skills in officiating, using AI to support feedback while maintaining deterministic logic for assessment.

The author states that referee education often measures whether someone remembers a law or reaches the final decision, but in real matches, the difficult skill is identifying the observable fact that makes the law apply. CounterCall aims to make this reasoning visible and testable.

It is described as not a federation product or endorsement, indicating it is an independent project with no institutional backing.

Inference: The positioning is that of a novel training method for officials, using AI-enhanced counterfactuals to improve decision-making. It is not positioned as a commercial product but as a proof-of-concept tool.

The description states the primary user is football referees, specifically those who need to train in identifying decisive observable facts and applying principles across similar cases.

It also mentions that the learning design can apply beyond football, to areas like handball offences, offside involvement, advantage, and disciplinary thresholds — suggesting a broader potential ICP.

Inference: The core ICP is football referees, with potential expansion into other sports or domains requiring principled decision-making. No evidence of specific customer segments or adoption data is provided.

The description does not mention any pricing, monetization, or business model. It is described as a self-built project submitted to a hackathon and not affiliated with any commercial entity.

Inference: There is no evidence of a business model or pricing structure. The tool appears to be experimental in nature.

The product is built using:

• Cloudflare Worker-compatible deployment
• Node.js server
• JavaScript, OpenAI Responses API (GPT-5.6 Sol)
• Codex for shaping and implementation
• A deterministic scenario engine for correctness and mastery

It uses structured outputs to ensure the model does not change grading. If the API fails or is unavailable, a deterministic fallback is used.

Inference: The technical stack suggests a lightweight, scalable architecture with AI integration for coaching and deterministic logic for assessment. The separation of roles between AI and logic is a key design choice.

The description states that the project was built independently in Moldova, submitted to a hackathon, and is not a federation product or endorsement.

There is no evidence of customers, revenue, usage metrics, or adoption. The tool is described as an experimental prototype.

Inference: No traction or maturity signals are evident. It is a self-reported prototype with no external validation or user base.

The description does not mention any competitors or existing tools in the space of referee training or decision-making coaching.

It implies that current methods focus on remembering laws or final decisions, rather than reasoning through observable facts — suggesting a gap in the market for this approach.

Inference: There is no evidence of direct competition. The tool may be addressing an unmet need in officiating education, but no existing solutions are referenced.

• No commercial traction or adoption: The tool is described as a hackathon submission with no evidence of real-world use.
• Dependency on AI API: If the OpenAI API fails or becomes unavailable, the experience may degrade.
• Limited scope: It is currently focused on football and lacks evidence of broader application or scalability.
• No institutional support: The lack of federation endorsement raises questions about credibility and potential for adoption.

Inference: The project is experimental and not yet proven in a commercial or educational setting. Risks include technical fragility, limited market relevance, and lack of validation.

1. What is the intended path to scaling this beyond a hackathon prototype?
2. How would you validate that learners are actually improving their reasoning skills through this method?
3. Are there any existing partnerships or pilot programs with football organizations?
4. What are the technical and operational risks of relying on GPT-5.6 Sol for coaching?
5. How do you plan to transition from a self-built tool to a sustainable product?